Aplikasi Pencari Info Obat dengan Masukan Citra Teks Kemasan Obat Berbasis Android Menggunakan Tesseract OCR Engine

Abstract

Detailed information on drug is usually written on a leaflet that can only be read when the package has been opened. It is difficult for consumers who just want to know the detailed information of certain drug to decide whether the drug will be bought or not. A solution is needed to convert drug brand text image into editable text, so that it can be used as a keyword query for drug info searching on the internet. Optical Character Recognition (OCR) is a process to convert image into editable text. Tesseract performs several processes to identify drug brand text image. The processes are image preprocessing, feature extraction, segmentation, and word recognition. This research developed an Android application to recognize drug brand text image. Tesseract is used because it has good accuracy in recognizing text with serif and sans serif font which have far density between characters. Drug packaging input image of the application can be taken from a camera or gallery. The best accuracy of this application in recognizing characters with 70 drug packaging images reached 96.80%.